Wikipedia talk:Manual of Style (dates and numbers)/Archive D3

What about date ranges?
We're guided to wikilink complete dates, for user date formatting preferences, like so: 14 November 2006. But what about 14-26 November 2006? &mdash; SMcCandlish &#91;talk&#93; &#91;contrib&#93; ツ 18:33, 12 February 2007 (UTC)


 * This is a big problem. There seems to be no way to do this without messing up someone's date preferences, other than 14 November 2006 – 26 November 2006 . Stephen Turner (Talk) 19:01, 12 February 2007 (UTC)


 * Date preferences don't apply for most readers (as they have no accounts) so it's only a problem for those who have accounts, and have prefs working the other way. Most readers will see date ranges according to the way they are written, and they will appear correctly. I've been treating date ranges by applying the correct format for the article.


 * Of course, this would be another nice thing to have if we are ever going to get rid of linked dates - date ranges formatted correctly according to prefs. --Pete


 * Actually, date preference linking does have some effect, even for people who are not logged in, as far as the formatting goes (commas or the lack of them, for example).


 * I usually ignore those who insist on using formats that will never work right anyway when the date doesn't include a year, so 14 November–26 November 2006 (i.e., 14 November–26 November 2006) will have acceptable results for most users. Gene Nygaard 03:23, 14 February 2007 (UTC)


 * If you decide to ignore the effects for some people's preferences, it is better not to link the dates at all. That way it will look at least consistent (e.g. 14 November–26 November 2006). To me the above example looks like "14 November–2006-11-26", which is clearly not acceptable. &minus;Woodstone 08:18, 14 February 2007 (UTC)

By the way, there is a template, daterange, that is underused at the moment. --Kevinkor2 15:38, 20 March 2007 (UTC)


 * What's the point of that template? It just turns into "D1 to D2". Stephen Turner (Talk) 16:19, 20 March 2007 (UTC)


 * daterange has a minor point: It highlights date ranges for semantic purposes. I suggest that if the date range problem has a technical solution, we rewrite daterange to incorporate it. --Kevinkor2 10:35, 21 March 2007 (UTC)

UTC± Notation
If the notation for a time range five hours ahead of UTC is UTC+5, then shouldn't the notation for five hours behind be UTC−5, not UTC-5? It seems like it should be a minus, not a hyphen, for consistency. 155.33.61.98 20:02, 14 February 2007 (UTC)
 * Honestly, does anyone in the world care? I'd hazard a guess that 99.999% of computer users around the world use the hyphen and are not even aware that the magic "true minus" Unicode character even exists. &mdash;  SMcCandlish &#91;talk&#93; &#91;contrib&#93; ツ 18:48, 16 March 2007 (UTC)
 * I care. The characters exist, and they exist for a reason, so why not use them when it's appropriate to do so? They also do indeed format differently, so if it's appropriate to use them in this context, I say use them. People don't care about a great many Wikipedia guidelines, but that doesn't mean they shouldn't exist.  I notice you use an MDASH before your signatures.  Why not a double–HYPHEN-MINUS?  "Who cares?"  You do.  &mdash; Fastolfe00 13:50, 8 April 2007 (UTC)


 * There is a good reason to use the minus entity, because it is non-breaking. So by using it, one avoids that the time zone designation ends up on the next line, separated from the negative sign. &minus;Woodstone 14:52, 8 April 2007 (UTC)

Question
Are you always supposed to link dates? This page doesn't seem to address that. – Lantoka  (talk) 23:50, 17 February 2007 (UTC)


 * See above for discussion on this point, but the short answer is that if you wikilink full dates, then they are formatted correctly to user preference. --Pete 10:04, 18 February 2007 (UTC)


 * I'm not sure why you say the page doesn't address that. It seems to me to discuss it at great length. In particular, see the sections Dates containing a month and a day and Partial dates. Stephen Turner (Talk) 10:10, 18 February 2007 (UTC)


 * Ah, I see. Took me a minute to puzzle through that one. Those two sections make it clear that if you do link you should do it their way so that user date preferences work, but don't explicitly state that dates should be linked all the time. Thus my confusion. Thanks for the responses. – Lantoka  (talk) 08:02, 19 February 2007 (UTC)
 * Indeed there are many people, including myself, who think that dates are overlinked, and have requested a wiki extension to allow (but not require) date preference formatting without linking. DES (talk) 15:06, 8 April 2007 (UTC)

Clarification
The section on birth/death dates does not explicitly cover the situation where the date of birth is fixed, but while death is certain, the exact date is not known. For example, consider Amelia Earhart and Fred Noonan. They were last seen on 2 July 1937 en route to Howland Island in the western Pacific. There is no concrete evidence regarding their demise. They may have died in a crash, reached land somewhere and survived briefly as cast-aways, etc. However, by now they are certainly dead. Consider Fred Noonan: what is well-documented are
 * the date he was born 4 April 1938,
 * the date went missing, 2 July 1937, and
 * the date he was declared dead by a court, 20 June 1938.

What should the correct format for birth/death dates be in this case? For the moment, it is listed as (4 April 1893-c.1937). There has also been a suggestion that this could be something like (4 April 1893-missing 2 July 1937, declared dead 20 June 1938). Any input?

Ronnotel 15:30, 21 February 2007 (UTC)


 * My view is that this is sufficiently uncommon that we don't need guidance about it in the Manual of Style; a solution appropriate to each individual article can be decided on a case-by-case basis on that article's talk page. Stephen Turner (Talk) 16:09, 21 February 2007 (UTC)
 * Agreed, and in this case I would list the death as "ca. 2 July 1937" because it is very probable (according to sources on the topic, not just my personal estimation). If it were less probable: An explorer took off into the Amazon on 2 July 1937 never to be seen again except perhaps by a hungry jaguar or some hostile Yanomamo, I would say "ca. July 1937" (or "ca. July 1937" for those who like to wikilink years without day & month) or even "ca. 1937" or "ca. 1937-40", depending upon the sourced (not original research/opinion) likelihood of his extended survival.  If it were even less probable: A young girl was abducted 2 July 1937 by a known sex-offender with a history of keeping captives for extended periods of time, but the girl is known for a fact to be deceased because her skull was found 20 years later and ID'd from dental records, but too damaged to determine age (or whatever; everything I know about forensics I know from watching CSI), then "some time after 1 July 1937". &mdash;  SMcCandlish &#91;talk&#93; &#91;contrib&#93; ツ 19:03, 16 March 2007 (UTC)

Proposed
When the biographical subject went missing and a year of death is not established in the historical record:


 * Amelia Earhart
 * (24 July 1897 – missing 2 July 1937, declared dead January 5 1939)

Gwen Gale 02:54, 25 February 2007 (UTC)


 * I already replied to this above, saying:
 * My view is that this is sufficiently uncommon that we don't need guidance about it in the Manual of Style; a solution appropriate to each individual article can be decided on a case-by-case basis on that article's talk page.
 * I've reverted the change you made on the main page. I apologise for saying in my comment that you'd raised this issue before and ignored the answer, because I now realise that it was someone else. Nevertheless, you should get a consensus on the talk page before changing the Manual of Style. Stephen Turner (Talk) 10:04, 26 February 2007 (UTC)


 * I hadn't received any comments at all. Do you have some sort of special authority on this page? Please clarify, thanks. By the way, uncommon or not, the lack of policy on this has caused wasted time and gnashing of teeth. A policy is needed (there are a few ways of handling this that would be ok with me, but a policy is needed). Gwen Gale 10:16, 26 February 2007 (UTC)


 * No, I don't have any special authority on this page. However, this exact question was raised by User:Ronnotel just a few days ago, and the only person who replied (me) said it wasn't necessary. So there certainly isn't a consensus for the change at the moment.
 * Personally, I realise that a group of editors has been struggling with this issue, but I still think that the Manual of Style is not the place to resolve issues like this. It should only provide guidance for situations which come up in many articles. It can't hope to document every variant which could come up in rare situations.
 * Maybe some more people would like to comment?
 * Stephen Turner (Talk) 10:32, 26 February 2007 (UTC)


 * I think your reasoning may be arbitrary. What quantitative threshold are you referring to? Please provide verifiable numbers if possible. Also, a reference to written WP policy supporting your statement as to this kind of threshold and its relevance to this discussion would be helpful. Thanks. Gwen Gale 10:41, 26 February 2007 (UTC)

Would anyone else like to comment on this? Or are you all too busy arguing about binary prefixes? :-) Stephen Turner (Talk) 10:25, 27 February 2007 (UTC)


 * Sounds like instruction creep to me. The case is rare enough that its specification would just clutter this MoS page, and should be decided on case-by-case basis. Either form is OK, and it's really not something worth arguing. Duja ► 10:41, 27 February 2007 (UTC)


 * Given I'm ok with a consensus I don't agree with, let's flip it: Is it fair to say that this project page provides no guidance on missing-and-maybe-later-declared-dead bios? If it does provide guidance, what is it? For example, can clarity be considered the goal, guidelined or not? Gwen Gale 10:53, 27 February 2007 (UTC)


 * I would think that "(24 July 1897 – missing 2 July 1937)" would be satisfactory for notational usage; the article itself can capture the legal tying-up of loose ends. Imagine, if you will, that her body was successfully (and incontrovertibly) identified today. We still wouldn't know the exact date of death, but "(24 July 1897 – missing 2 July 1937, declared dead 5 January 1939, confirmed dead 1 March 2007)" starts becoming a lot of baggage to carry around. Askari Mark (Talk) 18:22, 1 March 2007 (UTC)

Manual of Style (dates and numbers)

 * "new users and unregistered users do not have any date preferences set, and will therefore see the unconverted ISO 8601 date."

I believe that excerpt is outdated since the software will convert the date to the standard European format (ie 2007-03-04 -> 4 March 2007 (default)) regardless of any settings. If people choose to use the US format, they would be using that appearance or whatever format they are comfortable with.

I think it is a better idea to promote the ISO format as it is a win-win situation.

-- Cat Chi? 22:02, 3 March 2007 (UTC)
 * I suggest that you turn off your date preferences, take a look at your example above, and draft a response here. --Pete 01:36, 4 March 2007 (UTC)


 * Even better, log out rather than turning off your preferences. It's not true, unfortunately. Stephen Turner (Talk) 09:57, 4 March 2007 (UTC)


 * This is something the software can (and should) do rather easily, based on the information sent by the browser or possibly the IP address of the anonymous user. I prefer ISO dates as being the only sensible format (large &rarr; small) and do not like that its use is being discouraged for stylistic reasons.  --Sapphic 22:22, 2 April 2007 (UTC)


 * You seem to be arguing that the software could convert ISO dates to normal text dates. Which it could, and that would be nice, but it's not relevant to us on this page. In the Manual of Style, we have to give style guidelines based on the software which we have available, and the current state of the software is that anonymous and new users will see ISO dates which will be unfamiliar to most of them. Stephen Turner (Talk) 08:23, 3 April 2007 (UTC)


 * Strange it was working before. -- Cat chi? 23:13, 2 April 2007 (UTC)
 * Sorry to have edited your contribution to a talk page, I didn't mean to be rude, but I thought maybe adding &lt;nowiki> stuff wasn't refactoring in a too terribly impolite way. My date viewing preferences made both of the items in your example appear identical, which I don't think is how you intended for them to be viewed.  --Sapphic 00:30, 3 April 2007 (UTC)

AD/BC CE/BCE issues
Without going through every single archived page above, I do see an early discussion about use of AD/BC and CE/BCE discussion in which took place in July 2004. Could someone steer me to a more recent archived discussion on this issue (if such a discussion has in fact taken place)? Spamreporter1 20:19, 10 March 2007 (UTC)
 * Well, the passage of over 24 hours since my post above does suggest that the AD/BC - CE/BCE issue may not have been re-visited since 2004. So . . . second request - Has there been a more recent discussion of the AD/BC - CE/BCE issue that anyone is aware of?  If so, could you please post the location of that archived discussion here. Spamreporter1 21:26, 11 March 2007 (UTC)
 * Some Google searching turns up: a discussion from May 2005, a proposal from June 2005, a discussion from August 2005, a discussion from December 2005, a discussion from January/February 2006, a discussion from September/October 2006, and dozens of discussions spanning May 2005 to November 2006 on Wikipedia talk:Eras. Incidentally, demanding people to do your Google searching for you within a 24-hour deadline is a bit presumptuous. --Delirium 21:42, 11 March 2007 (UTC)
 * Kind of you, and apologies - your search skills are superior to mine - I didn't realize that Google could be used to search archived pages. Thanks and apologies again. Spamreporter1 21:47, 11 March 2007 (UTC)
 * I was a little bit upset by the CE/BCE usage and argued on this subject in the CE talkpage in 3 consecutive sections, here is a link to the first: Talk:Common_Era, Magnus Andersson (talk) 12:03, 21 June 2008 (UTC)

Ranges of years
The only reference I can find to using ranges of years is where the manual says, "Do not use two digits to express a year unless at the end of a range, e.g., "1970–87" (the same for BC)."

So, it's acceptable to use the two-digit format in a range of years. But, is it preferable? Personally I find 1970–87 scans much beter than 1970–1987, and is just as easy to understand. But I've seen people changing it when I've used it. From the way the guidelines are currently written, neither is recommended, so should the general practice of "not changing format for no good reason" apply? – Kieran T  (' talk ') 15:50, 19 March 2007 (UTC)
 * It's most common and probably best to have full dates. —Centrx→talk &bull; 21:12, 19 March 2007 (UTC)


 * I don't feel strongly either way, but have to observe that it is "most common" only on Wikipedia and rather unusual in print, and that the WP preference for this almost certainly derives from some editors' mania for wikilinking years even when the MoS discourages the practice. &mdash; SMcCandlish &#91;talk&#93; &#91;contrib&#93; ツ 21:24, 19 March 2007 (UTC)


 * I'd agree that it seems more common to use the shorter form in print. And that's an interesting point about the wiki-linking years mania. But for balance, Centrx, could you please expand on why you think it's better to use them in full? Is the other side of this coin, perhaps a sense that "abbreviation" looks lazy? – Kieran T  (' talk ') 22:13, 19 March 2007 (UTC)
 * And just to be clear, I would add that I don't think Centrx is a mad overlinker of years; his observation, if limited to wikitext, is correct. I'm just explicitly stating why it is correct (from what I can determine; others may have other theories), because that might bear on whether the don't-abbreviate trend actually has a useful basis one should pay much heed to. That's all. &mdash;  SMcCandlish &#91;talk&#93; &#91;contrib&#93; ツ 23:11, 19 March 2007 (UTC)


 * It doesn't say that the abbreviated form is preferable, and it should stay that way. I think we should specify that one digit abbreviations of years (1983-7) are unacceptable, however, and should be changed whenever we run across them.  Gene Nygaard 04:20, 20 March 2007 (UTC)


 * I'd be (sincerely) grateful if people would give reasons as well as opinions for this, because otherwise we just have a vote rather than a discussion. Your suggestion that it should "stay that way" leaves me wondering, do you think therefore that it should state that we "prefer" the full version (and if so, on what basis?), or are you saying the choice/ambiguity should remain? – Kieran T  (' talk ') 13:01, 20 March 2007 (UTC)


 * Ambiguity isn't necessarily a bad thing. There isn't and isn't likely to be soon any consensus on systemwide changes in this regard.  We don't need to specify everything.  Settle for achieving consistency within an article; in a great many cases that in itself goes a long ways towards alleviating any potential problems.  Gene Nygaard 00:28, 21 March 2007 (UTC)


 * Sounds reasonable. &mdash; SMcCandlish &#91;talk&#93; &#91;contrib&#93; ツ 00:11, 21 March 2007 (UTC)
 * I don't see any basis for saying that "1983-87" is "acceptable" but "1983-7" isn't. If anything, the latter is far more common. &mdash;  SMcCandlish &#91;talk&#93; &#91;contrib&#93; ツ 00:11, 21 March 2007 (UTC)


 * I doubt very much that it is more common. Certainly not in what I see outside Wikipedia.  We normally deal with abbreviated years in two-digit manner, if we are going to use some short form to express a date, for example.  It doesn't make any sense whatsoever to allow anything other than that two-digit shortening when expressing date ranges.  Gene Nygaard 00:28, 21 March 2007 (UTC)
 * But why? I'm not seeing what the rationale is. &mdash;  SMcCandlish &#91;talk&#93; &#91;contrib&#93; ツ 19:35, 21 March 2007 (UTC)
 * Anecdotally (we all know what THAT counts for), it seems that people will (in speech) more commonly say things like "I was in college from nineteen ninety-one to ninety-five" than they will "nineteen ninety-one to five". I'm not certain that's on point to this issue anyway. --Dante Alighieri | Talk 22:57, 21 March 2007 (UTC)
 * I agree with Gene Nygaard in that I see the likes of "1983-87" far more often than "1983-7". I personally prefer using "1983-1987", but have no real problem with "1983-87".  However, I’ve always found the style "1983-7" to be irksome; I always feel uncomfortable that there might be an error.  Consider a different example like "1926-7":  could "1926-37" have been meant and what I’m seeing is a typo?  There are occasions when the context might not make it clear.  That’s the best reason I can offer.  Askari Mark (Talk) 03:47, 26 March 2007 (UTC)

Now it says to use four digits except for years after 999 AD. I try to change to "all digits", which would work all the time, but some disagree. What about year 1,000,000? Not happen yet, you say? 1,000,000 BCE!--MajorHazard 11:30, 7 April 2007 (UTC)

When did the year-without-date recommendations change?
WP:MOSDATE used to recommend that years without dates should not be linked. When did this change, and why? I still regularly see editors removing such links "per WP:MOSDATE". —Josiah Rowe (talk • contribs) 17:41, 23 March 2007 (UTC)


 * It was last May. See /Archive 48 for the discussion. But in summary, the previous version had gradually become stronger over the years, didn't command consensus, and was the subject of endless arguments. Stephen Turner (Talk) 21:09, 23 March 2007 (UTC)


 * Whaddya know. I wonder what other guidelines have changed unbeknownst to editors who are still editing by an out-of-date version. Ah, well — thanks for the history pointer. —Josiah Rowe (talk • contribs) 22:05, 25 March 2007 (UTC)

The guideline states that dates should be linked only in accordance with WP:MOSLINKS, just like any other link. The only inconsistency is that all the example dates are linked for no good reason. —Centrx→talk &bull; 22:08, 25 March 2007 (UTC)


 * That's pretty amusing (though of course also ultimately confusing and unhelpful). Who wants to fix it? PS: I refer to the problem in the guideline, not to Centrx's comment on it, of course. &mdash;  SMcCandlish &#91;talk&#93; &#91;contrib&#93; ツ 00:14, 26 March 2007 (UTC)